SURFACE ACOUSTIC WAVE STUDIES OF DEFECTS IN ELECTRON IRRADIATED GaAs

HAL Id: jpa-00225323

https://hal.archives-ouvertes.fr/jpa-00225323

Submitted on 1 Jan 1985

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

SURFACE ACOUSTIC WAVE STUDIES OF

DEFECTS IN ELECTRON IRRADIATED GaAs

M. Brophy, A. Granato

To cite this version:

M. Brophy, A. Granato.

SURFACE ACOUSTIC WAVE STUDIES OF DEFECTS IN

JOURNAL DE PHYSIQUE

Colloque CIO, supplément au n"12, Tome 46, décembre 1985

page

C10-541

SURFACE ACOUSTIC WAVE STUDIES OF DEFECTS I N ELECTRON IRRADIATED GaAs

M.J. BROPHY'AND A.V. GRANATO

Dept. of Physics and Materiaïs Research Laboratory, University

of Illinois, Urbana, Il. 61801, U.S.A.

Abstract Surface Acoustic Waves (SAW) provide a means of studying low temperature anelastic relaxations of crystal defects produced by irradiation. Due to their confinement near the surface where radiation induced defect concentrations are highest. SAW are more sensitive than bulk ultrasonic waves in these investigations. Further- more. for piezoelectric materials. the use of photolithographically deposited interdigital SAW transducers eliminates the ubiquitous ultrasonic bond problem faced in bulk wave experiments, facilitating annealing studies. Results are reported for 2.5 MeV electron irradiated GaAs. Two peaks in attenuation versus temperature were observed both with SAW and with bulk waves.

1. Introduction

In the usual ultrasonic pulse-echo studies of anelastic relaxations of point defects. the role of the ultrasonic pulse is to provide an external perturbation which biases the defect configuration energies. This leads to a thermal repopulation of the energy levels with a characteristic relaxation time, and thus to the usual Debye attenuation peak and velocity dispersion.

This role can be performed by any kind of stress wave assuming the displacements involved are such as to produce the biasing effect. In what follows we discuss the use of Rayleigh surface acoustic waves (SAW) to provide that bias strain. and present results for defects in electron irradiated GaAs.

There are two displacements in a Rayleigh wave. One is longitudinal and the other is transverse. per- pendicular to the surface. Thus the particle motion is elliptical. Since the displacements Vary differently with depth. detailed particle.motion is also a function of depth. Displacement amplitude versus depth for both displacements for a

<

110> propagating SAW on a (001) plane of GaAs is shown in Fig. 1.'

SAW can be generated in a number of ways.' Al1 electronic SAW devices are fabricated on piezoelec- tric substrates using interdigital transducers3 deposited photolithographically in the proper orientation. In this method, two interleaved comb structures are laid on the surface and an rf voltage applied between them (Fig. 2 ) . This induces alternating strains which produces a Rayleigh wave with a wavelength equal to twice the spacing between adjacent electrodes ("fingers").

The most significant advantage of using SAW compared to using bulk ultrasonic waves is that with SAW we do not have the problem of making an ultrasonic bond that will survive to cryogenic tempera- tures. Moreover. for radiation damage annealing experiments we need a bond that will survive anneals around room temperature and then revert to its original form when cooled. We found no bonding agent which could satisfy both these requirements for GaAs. On the other hand the SAW devices were run from 4.2 K to 330 K. and the measured attenuation reproduced extremely well.

(1) Present Address - Coordinated Science Laboratory. Univ. of Illinois and Electronic Decisions Inc.. Ur- bana. IL 61801. USA

CIO-542

JOURNAL

DE

PHYSIQUE

Another advantage in using SAW is that a thin sample is used. so the damage is relatively uniform. The thick bulk samples have nonuniform damage in the damaged region and no damage at al1 in the rest of the sample. A further advantage of thin samples is that heating by the beam is minimized. This would allow irradiation a t very low temperatures without having to go to a miniscule beam current.

Insulating samples can only be done with SAW. In fact. SAW transducers require an insulating sub- strate. Irradiation of a bulk insulating sample would lead to a build up of charge on the sample which could lead to dislocation formation. In this case the two methods are complementary.

The major disadvantage of SAW is the complicated form of the wave. As mentioned earlier. the SAW has both longitudinal and transverse components whose magnitudes Vary differently with depth. This is to be compared with the simple plane wave form and simple polarization of the bulk waves. Also. in GaAs a strong SAW can only be produced on a (100) plane in

a

[Oll] direction. Any other allowed SAW is very weak. As a result the determination of defect symmetry with SAW is a problem, whereas it is straightforward in the bulk wave measurements.

The fabrication of a SAW device is also much more difficult than the set-up required for a bulk wave expriment. One must be trained in photolithographic techniques and in the use of microfabrication e q u i p ment.

In summary. the ideal arrangement would be to use both bulk and surface acoustic waves to take advantage of the special attributes of each. This is what we have done in this work.

2. Experimental Method

Our goal was t o use SAW with an ultrasonic pulse echo apparatus. Thus we needed to design an effective SAW reflector to obtain the usual series of echoes. and we needed to be able to produce sizabie SAW pulses in Our samples. GaAs is weakly piezoelectric. so that generation of SAW with interdigital transducers on GaAs is well e ~ t a b l i s h e d . ~ The strongest waves are generated in [Il01 directions on (100)

planes.

The design criteria for SAW reflectors are well k n o ~ n . ~ The contributions to the reflection come from mechanical and piezoelectric sources. The mechanical contribution is due to the mass and elasticity of the reflecting electrodes, and to a lesser extent due to the substrate distortion caused by the electrode. The piezoelectric sources of reflection are basically a regeneration of a SAW as the incident SAW passes under the electrodes. Since GaAs is only weakly piezoelectric. the mechanical effects dominate. Briefly stated. good reflection should be obtained with many thick reflecting electrodes. This was indeed found to be the case.

I t was decided to make reflectors by having 2 identical IDT's with a 5 mm space between them. Obviously f o r multiple echoes the generating transducer itself must be a good reflector. and by having two IDT's instead of one IDT and one reflecting array the yield of usable devices should be improved. If one IDT was shorted or otherwise defective the device was still operable provided the other IDT was good. Two pairs of IDT's a t right angles to one another were used to allow us to do two frequencies a t once. Obviously their propagation directions are equivalent on a (100) plane of GaAs. a cubic crystal.

The SAW devices were fabricated using standard lift-off photolithographic

technique^.^-^

A typical device is shown in Fig. 2. In the radiation damage experiments. only the 5 mm diameter area between the IDT's was irradiated.

The bulk ultrasonic measurements were done using 3-Methyl pentane. applied a t about 150

K.

as a bonding agent. These bonds worked well below about 80 K.

The samples used in this work were al1 single crystal GaAs. The SAW wafer was Semi-Insulating

LEC grown GaAs. and the bulk material was n-type Horizontal Bridgman grown material.

Al1 irradiations were carried out with 2.5 MeV electrons. with the sample temperature never exceed- ing 190 K. Experimental details will be reported elsewhere in a fuller account of this work.

3. Experimental Results

Fig. 3 shows the measured change in logarithmic decrement versus temperature for electron irradiated GaAs using both the Cu bulk waves and SAW. The SAW sample electron dose has been normalized to the bulk sample dose of 8.5 X l O " ~ m - ~ . Two distinct defects which undergo anelastic relaxation are seen in both cases. No peak occurs for the C' bulk mode indicating that these defects possess trigonal symmetry.

plane of GaAs has a transverse component which is equivalent to a C,, mode displacement. Moreover. in this case the "C,," component of the SAW is much larger than the longitudinal component.

The second observation from Fig. 3 is that the SAW peaks are significantly bigger than the buik peaks for the same dose. Part of the difference can be explained by what percent of the volume traversed by the wave is damaged by the irradiation.

In the SAW device (Fig. 2) almost al1 of the region between the transducers is irradiated. Further- more we see from Fig. 1 that the SAW displacements are almost entirely confined to a depth of one wavelength (20 p m for a 145 MHz SAW). In this region near the surface the expected damage rate is high and can ôe assumed to be uniform.

On t h e other hand t h e situation is considerably l e s favorable for defect detection in the buik case. The ultrasonic beam diameter, which. neglecting diffraction. equals the transducer diameter. is uniformly irradiated. However. the ultrasonic pulse traverses the whole length of the sample but only part of this length is damaged. In this case the bulk sample was 12.2 mm thick. We expect the damaged region to extend about 2 mm into the sample. so only about 16% of the sample was damaged, i.e. contained the defects responsible for the peak.

Furthermore. in the damaged volume of the bulk sample the damage is not uniform with depth. If we distribute this damage uniformly with depth w e are left with an effective damage rate of about half the rate a t t h e surface.

We cannot express the SAW relaxation strength in any simple way for comparison with that of Cd,. No theory is available for the problem of SAW attenuation in a medium which is lossy in different ways for the t w o component displacement amplitudes (Fig. 1).

The problem is not equivalent to that of shear and longitudinal bulk waves propagating in the same direction. In that case the attenuations of the t w o modes are independent. In the SAW case the two dis- placements are coupled. so that changes in one mode's displacements will cause corresponding changes in the other mode. This must occur so that the SAW is a t al1 times a solution to the appropriate wave equa- tion. i.e. of the form shown in Fig. 1.

4. Conclusions

We have seen that Surface Acoustic Waves can be used to study anelastic relaxations of point defects in crystals.

The fact that SAW studies of this kind are possible suggests many possible applications. Radiation damage in thin epitaxial layers could be examined ultrasonically. Damage by less penetrating radiation. for example ion implantation damage before and after annealing. could be investigated. Of course. other piezoelectric materials which can support SAW could be studied as well.

The use of SAW in defect studies could prove to be a powerful new tool for many applications. Acknowledgements

This work was partially supported by the Department of Energy. Division of Materials Sciences. under contract DE-AC02-76ER01198.

Ref erences

1. These plots were obtained from a program developed by Dr. Supriyo Datta for Professor Hunsinger's SAW group a t the University of lllinois a t Urbana-Champaign. The method of calculation is described in S. Datta and B. Hunsinger. J. Appl. Phys. 49, 2, 475 (1978).

2. G.W. Farnell. "Physical Acoustics," Vol. 6. p. 109, W.P. Mason and R.N. Thurston, Ed.. Academic Press, New York (1970).

3. A.A. Olliner. Ed., "Acoustic Surface Waves," Springer Verlag. Berlin (1978).

4. T.W. Grudkowski, G.K. Montress, M. Gilden and J.F. Black. Proc. 1980 IEEE Ultrasonics Symposium. p. 88 (1980).

CIO-544

JOURNAL DE PHYSIQUE

6. H.1. Smith, F.J. Bachner and N. Efrernnw, J . Electrochem. Soc. 1 1 , 5. 821 (1971

1.

7. R.W. Halverson. M.W. Maclntyre. W.T. Motsiff. IBM 3. Res. Develop. 26.5, 590 (1 982). 8. G.W. Collins, C.W. Halstead, IBM J. Res. Develop. 26, 5.596 (1982).

(001) Cut GaAs - (1 10) Propogating SAW transverse<001> PA= l mW/X

-

- - - 0.0 0.5 I .O 1.5 2 .O Depth (wavelengths)

Fig. 1 Displacemenls for SAW in GaAs.

SAW Resonator Confisuration Vertical IDT'S Horizontal IDT'S Finger Width, w

k m )

IO 5 Approximate Center 71.5 143 Frequency (MHz)

Colculoted Reflect ion .95 .99+ Coefficient

Both IDT Pairs Duty Cycle = 50% Wovelength = 4w Fingerlength = 105w Beam Width = IOOw Number of Finger Pairs = 150 Metallization: 25501 Al on 660A Cr

Fig. 2. Typical SAW device. Fig. 3. Logarithmic decrement v s T for 2.5